In this paper, we present an innovative framework whose objective is to allow drivers to recharge their Electric Vehicles (EVs) from the most environmentally friendly chargers using an intelligent hoarding approach. These chargers maximize renewable (e.g., solar) self-consumption, minimizing this way CO2 production and also the need for expensive stationary batteries on the electricity grid to store renewable energy that cannot be used otherwise. We model our problem as a Continuous k-Nearest Neighbor query, where the distance function is computed using Estimated Components (ECs), i.e., a query we term CkNN-EC. An EC defines a function that can have a fuzzy value based on some estimates. Specific ECs used in this work are: (i) the (available clean) power at the charger, which depends on the estimated weather; (ii) the charger availability, which depends on the estimated busy timetables that show when the charger is crowded; and (iii) the derouting cost, which is the time to reach the charger depending on estimated traffic. We devise the EcoCharge framework that combines these multiple non-conflicting objectives into an optimization task providing user-defined ranking means through an intuitive mobile GIS application. Particularly, our core algorithm uses lower and upper values derived from the ECs to recommend the top ranked EV chargers and present them through an intuitive map user interface to users. Our experimental evaluation with extensive synthetic and real traces from Germany, China, and USA along with EV charger data from Plugshare shows that EcoCharge meets the objective functions in an efficient manner, allowing continuous recomputation on the edge devices (e.g., Android Automotive OS, Android Auto or Apple Carplay).

In this demonstration paper, we present an innovative framework for sustainable Electric Vehicles (EVs) charging, dubbed EcoCharge, which utilizes an intelligent energy hoarding approach. Particularly, EcoCharge employs a Continuous k-Nearest Neighbor query, where the distance function is computed using Estimated Components (ECs) (i.e., a query we term CkNN-EC). An EC defines a function that can have a fuzzy value based on some estimates. Specific ECs used in this work are: (i) the (available clean) power at the charger, which depends on the estimated weather; (ii) the charger availability, which depends on the estimated busy timetables that show when the charger is crowded; and (iii) the derouting cost, which is the time to reach the charger depending on estimated traffic. Our framework combines these multiple non-conflicting objectives into an optimization task providing user-defined ranking means through an intuitive spatial application. The algorithm utilizes lower and upper interval values derived from ECs to recommend the top ranked EV chargers and present them through a map interface to users. We demonstrate EcoCharge using a complete prototype system developed using the Leaflet - OpenStreetMap library. In our demonstration scenario, attendees will have the opportunity to observe through mobile devices the benefits of EcoCharge by simulating its execution over various scheduled trips with real data retrieved from API requests (i.e., ECs).

Energy efficiency has become a primary optimization objective due to the global energy crisis and high levels of CO2emissions. Climate and energy targets have been leading to a growing utilization of solar photovoltaic power generation in residential buildings. As the number of IoT devices drastically increases, their automation through an intelligent home energy management system can provide energy and peak demand savings. The planning optimization of devices can be very challenging due to the unsophisticated user-defined preference rules. Existing solutions face convergence difficulties due to the management of multiple IoT devices tackling multiobjective problems. In this paper, we propose an innovative IoT data system, coined GreenCap, which utilizes a Green Planning evolutionary algorithm for load shifting of IoT-enabled devices, considering the integration of renewable energy sources, multiple constraints, peak-demand times, and dynamic pricing. We have implemented a complete prototype system available on Raspberry Pi and linked with openHAB framework. Our experimental evaluation with extensive real traces shows that the GreenCap prototype system efficiently generates a sustainable plan obtaining high levels of user comfort 92-99% along with ≈52% of self-consumption, while reducing ≈35% of the imported energy from the grid and ≈40% of CO2emissions.

In this demonstration paper, we present an innovative IoT data platform, coined GreenCap, which utilizes a Green Planning evolutionary algorithm for load shifting of IoT-enabled devices in smart environments that feature renewable energy sources. Particularly, GreenCap deploys a hybrid genetic algorithm with domain-specific local search heuristics, which results in a Memetic Algorithm (MA) that offers users an energy efficient allocation plan of their IoT devices, based on their personal preference rules (e.g., operate AC from 10am - 1pm). Our system allocates operations in the daily time-slots considering devices’ energy bounds to minimize the imported energy from the grid, exploit self-consumption and maximize users’ comfort. We demonstrate GreenCap using a complete prototype system available on Raspberry Pi, developed in Laravel using MariaDB and linked to openHAB framework. In our demonstration scenario, attendees will be able to observe through mobile devices the benefits of GreenCap by simulating its execution with real data for one week, using pre-configured or custom rules.

This demo paper presents EnterCY, an integrated Virtual and Augmented Reality Tourism platform for Cyprus. The platform's web-based, spatio-temporal virtual exploration component allows potential visitors to explore the rich cultural heritage, variety of activities, and wealth of sightseeing locations in Cyprus before their visit. EnterCY also enhances tourists' experiences during their visit through its mobile component, which offers on-site visual and audio guidance, personalized recommendations, as well as entertaining and learning features (e.g., story-telling), based on mobile-friendly Augmented Reality, location-awareness and Machine Learning technologies. Through Immersive Reality technologies, the platform provides for an after visit experience by creating personalized 360 video mementos of tourists' tours and supports integrated features that allow for experience sharing in popular social media platforms.

Internet of Things (IoT) devices have penetrated massively into smart environments (e.g., smart-homes, smart-cars or more generally smart-anything). Besides data collection, many IoT devices also enable the execution of Rule Automation Workflows (RAW), which span from simple predicate statements to procedural workflows capturing a smart actuation pipeline. RAW aim to meet the convenience (comfort) level of users under specific conditions (e.g., raise room temperature to 22 C if cold), but unfortunately cannot express long-term objectives of users (e.g., consume less than 400 kWh in December). In this paper, we present an innovative system, coined IoT Meta-Control Firewall (IMCF), which internally deploys an AI-inspired Energy-Planner (EP) algorithm that exploits domain-specific operators to balance the trade-off between convenience and energy consumption in satisfying the RAW pipelines of users. IMCF filters the RAW pipelines in a way that these do not conflict with the long-term objectives of users (like a network firewall). Our experimental evaluation with extensive real traces from an apartment, a house, and campus dorms shows that IMCF achieves very high levels of user convenience while remaining within the target energy consumption budgets expressed by users.

The Internet of Things (IoT) revolution has introduced sensor-rich devices to an ever growing landscape of smart environments. A key component in the IoT scenarios of the future is the requirement to utilize a shared database that allows all participants to operate collaboratively, transparently, immutably, correctly and with performance guarantees. Blockchain databases have been proposed by the community to alleviate these challenges, however existing blockchain architectures suffer from performance issues. In this short paper we propose Triastore, a novel permissioned blockchain database system that carries out machine learning on the edge, abstracts machine learning models into primitive data blocks that are subsequently stored and retrieved from the blockchain. Triastore comprises of two internal routines, namely: (i) Proof of Federated Learning (PoFL), which trains in a distributed manner a global model for the ingested data; and (ii) Blockchain Consensus, which commits this generated model data on permissioned blockchain database. We present a detailed explanation of our data ingestion algorithm with relevant examples and carry out an experimental evaluation with image data from MNIST. The evaluation shows that our proposed data ingestion framework retains high levels of accuracy with low loss in data quality.

In this paper, we present a novel decaying operator for Telco Big Data (TBD), coined TBD-DP (Data Postdiction). Unlike data prediction, which aims to make a statement about the future value of some tuple, our formulated data postdiction term, aims to make a statement about the past value of some tuple, which does not exist anymore as it had to be deleted to free up disk space. TBD-DP relies on existing Machine Learning (ML) algorithms to abstract TBD into compact models that can be stored and queried when necessary. Our proposed TBD-DP operator has the following two conceptual phases: (i) in an offline phase, it utilizes a LSTM-based hierarchical ML algorithm to learn a tree of models (coined TBD-DP tree) over time and space; (ii) in an online phase, it uses the TBD-DP tree to recover data within a certain accuracy. In our experimental setup, we measure the efficiency of the proposed operator using a ~10GB anonymized real telco network trace and our experimental results in Tensorflow over HDFS are extremely encouraging as they show that TBD-DP saves an order of magnitude storage space while maintaining a high accuracy on the recovered data.

In this demonstration paper, we present the TBD-DP operator, which relies on existing Machine Learning (ML) algorithms to abstract Telco Big Data (TBD) into compact models that can be stored and queried when necessary. Our proposed TBD-DP operator has the following two conceptual phases: (i) in an offline phase, it utilizes a LSTM-based hierarchical ML algorithm to learn a tree of models (coined TBD-DP tree) over time and space; (ii) in an online phase, it uses the TBD-DP tree to recover data within a certain accuracy. Our framework also includes visual and declarative interfaces for a variety of telco-specific data exploration tasks. We demonstrate the efficiency of the proposed operator using SPATE, which is a novel TBD visual analytic architecture we have developed. Our demo will enable attendees to interactively explore synthetic antenna signal traces, we will provide, in both visual and SQL mode. In both cases, the performance of the propositions will be quantitatively conveyed to the attendees through dedicated dashboards.

Predominant smartphone OS localization subsystems currently rely on server-side localization processes, allowing the service provider to know the location of a user at all times. In this paper, we propose an innovative algorithm for protecting users from location tracking by the localization service, without hindering the provisioning of fine-grained location updates on a continuous basis. Our proposed Temporal Vector Map (TVM) algorithm, allows a user to accurately localize by exploiting a k-Anonymity Bloom (kAB) filter and a bestNeighbors generator of camouflaged localization requests, both of which are shown to be resilient to a variety of privacy attacks. We have evaluated our framework using a real prototype developed in Android and Hadoop HBase as well as realistic Wi-Fi traces scaling-up to several GBs. Our study reveals that TVM can offer fine-grained localization in approximately four orders of magnitude less energy and number of messages than competitive approaches.

Wi-Fi (or WLAN) based indoor navigation applications for mobiles rely on cloud-based services (s) that take care of a user's (u) localization task using structures called Radio Maps (RMs). It is imperative for u to have a stable WiFi connection in order to either continuously receive location updates from s or to download RMs a priori for offline navigation. Wi-Fi networks however, suffer from intermittent connectivity due to poor network planning that results in sparse deployment of access points and effectively areas where Wi-Fi coverage cannot be guaranteed. This inherently affects the localization accuracy and therefore the navigation experience of users. In this paper, we propose an innovative framework for accurate and fast indoor localization over an intermittently connected WiFi network, coined Prefetching Localization (PreLoc). In Preloc, we prioritize the download of RM records based on knowledge acquired from historic traces of other users inside the same building. Instead of downloading the complete RM from s to u, we propose a Probabilistic Group Selection (PGS) strategy, which identifies RM records that have a higher probability of being necessary to a user moving inside a target area. We have evaluated our framework using a real prototype developed in Android, as well as realistic Wi-Fi traces we collected at the University of Cyprus. Our experimental study reveals that PreLoc using PGS and conventional fingerprint-based indoor positioning algorithms can yield accuracy that is as good as using the same algorithms with a complete RM, even under scenarios of weak Wi-Fi coverage.

In this paper we present a comprehensive architecture to carry out experimental repeatability studies on clusters of smartphones. Our architecture is founded on Smart Lab, our in-house architecture for managing real and virtual smartphones via an intuitive Web user interface. Our presented architecture consists of several exciting components for re-programming and instrumenting smartphones to perform application testing and data gathering in a facile manner, as well as executing mockup experiments by "feeding" the devices with GPS/sensor readings. We will particularly demonstrate the various components of our architecture that encompasses smartphone sensor data collected by mobile users and organized in our distributed NoSQL document store. The given datasets can then be replayed on our test bed comprising of real and virtual smartphones accessible to developers through our Web 2.0 user interface. We present the applicability of our architecture through various mockup experiments over different application scenarios.

Indoor Positioning Systems (IPS) have recently received considerable attention, mainly because GPS is unavailable in indoor spaces and consumes considerable energy. On the other hand, predominant Smartphone OS localization subsystems currently rely on server-side localization processes, allowing the service provider to know the location of a user at all times. In this paper, we propose an innovative algorithm for protecting users from location tracking by the localization service, without hindering the provisioning of fine-grained location updates on a continuous basis. Our proposed Temporal Vector Map (TVM) algorithm, allows a user to accurately localize by exploiting a k-Anonymity Bloom (kAB) filter and a bestNeighbors generator of camouflaged localization requests, both of which are shown to be resilient to a variety of privacy attacks. We have evaluated our framework using a real prototype developed in Android and Hadoop HBase as well as realistic Wi-Fi traces scaling-up to several GBs. Our analytical evaluation and experimental study reveal that TVM is not vulnerable to attacks that traditionally compromise k-anonymity protection and indicate that TVM can offer fine-grained localization in approximately four orders of magnitude less energy and number of messages than competitive approaches.

In this demonstration paper, we present a novel framework for searching objects (e.g., images, videos, etc.) captured by the users in a mobile social community. Our framework, is founded on an in-situ data storage model, where captured objects remain local on their owners smartphones and searches then take place over a novel lookup structure we compute dynamically. Initially, a query user invokes a search to find an object of interest. Our structure concurrently optimizes several conflicting objectives (i.e., it minimizes energy consumption, minimizes search delay and maximizes query recall), using a Multi-Objective Optimization approach and calculates a diverse set of high quality non-dominated Query Routing Trees (QRTs), in a single run. The optimal set is then forwarded to the query user (decision maker) to select a particular QRT to be searched based on instant requirements and preferences. To demonstrate the capabilities of SmartP2P during the conference, we will utilize our cloud of smartphone devices, i.e. the SmartLab testbed composed of 40+ Android smartphones and tablets, as well as mobility and social patterns derived by Microsofts Geolife project, DBLP and Pics n Trails. We will allow the attendees to use a real SmartLab Android device to query our local Smartphone Network using any of the four algorithmic choices provided by the SmartP2P framework. The query device will then be provided with the optimal QRTs and the attendees will be able to visually decide the optimal QRT to be searched. A P2P search on the Smartphone Network will follow making available to the query user the desired objects of interest, in an optimal manner. The conference attendees will be able to appreciate how social content can be efficiently shared with other attendees within close proximity without revealing their personal content to a centralized authority.

The bulk of social network applications for smart phones (e.g., Twitter, Face book, Foursquare, etc.) currently rely on centralized or cloud-like architectures in order to carry out their data sharing and searching tasks. Unfortunately, the given model introduces both data-disclosure concerns (e.g., disclosing all captured media to a central entity) and performance concerns (e.g., consuming precious smart phone battery and bandwidth during content uploads). In this paper, we present a novel framework, coined Smart Opt, for searching objects (e.g., images, videos, etc.) captured by the users in a mobile social community. Our framework, is founded on an in-situ data storage model, where captured objects remain local on their owner's smart phones and searches then take place over a novel lookup structure we compute dynamically, coined the Multi-Objective Query Routing Tree (MO-QRT). Our structure concurrently optimizes several conflicting objectives (i.e., it minimizes energy consumption, minimizes search delay and maximizes query recall), using a Multi-objective Evolutionary Algorithm based on Decomposition (MOEA/D) that calculates a diverse set of high quality non-dominated solutions in a single run. We assess our ideas with mobility patterns derived by Microsoft's Geolife project and social patterns derived by DBLP. Our study reveals that Smart Opt can yield query recall rates of 95%, with one order of magnitude less time and two orders of magnitude less energy than its competitors.